The Solstice and Equinox systems, developed together with NVIDIA, Oracle and Argonne National Laboratory, will exceed the processing performance of El Capitan and Frontier, the current world leaders.
NVidia Headquarters PHOTO: Facebook
The project, announced on Tuesday, raises the bar in the world competition for supremacy in artificial intelligence. The initiative brings together NVIDIA, Oracle, Argonne National Laboratory and the United States Department of Energy, which are collaborating on the development of the Solstice and Equinox systems, considered the future world benchmarks of AI performance, according to Mediafax.
Supercomputers for research
Solstice will include more than 100,000 NVIDIA Blackwell GPUs, making it the largest AI supercomputer ever built in the US, while Equinox will integrate approximately 10,000 Blackwell GPUs. According to the American Nuclear Society, construction of the latter will begin immediately at the Argonne Leadership Computing Facility, with delivery planned for 2026.
The two systems will be connected to the Department of Energy’s national network of scientific instruments and data centers to support research in critical areas such as energy, security and basic sciences.
Winning the AI race
Paul Kearns, director of Argonne National Laboratory, says, “The Equinox and Solstice systems are designed to accelerate a broad set of AI workflows, and the collaboration with Oracle and NVIDIA will prepare thousands of researchers to harness the revolutionary capabilities of these technologies.”
“Winning the AI race requires new and creative partnerships that bring together America’s brightest minds and industries”said Energy Secretary Chris Wright. “The two systems will be a real engine of scientific and technological innovation”.
The US remains the world leader in AI
With performance estimated at more than 2 exaflops, Solstice and Equinox are expected to surpass El Capitan and Frontier, the most powerful supercomputers in the world at the moment, strengthening the US position as a world leader in artificial intelligence and high-performance computing.
